Received: by taz.hyperreal.com (8.6.12/8.6.5) id JAA24020; Sat, 3 Feb 1996 09:21:01 -0800 Received: from sierra.zyzzyva.com by taz.hyperreal.com (8.6.12/8.6.5) with ESMTP id JAA24008; Sat, 3 Feb 1996 09:20:55 -0800 Received: from zyzzyva.com (randy@localhost [127.0.0.1]) by sierra.zyzzyva.com (8.6.12/8.6.11) with ESMTP id LAA28205 for ; Sat, 3 Feb 1996 11:20:51 -0600 Message-Id: <199602031720.LAA28205@sierra.zyzzyva.com> To: new-httpd@hyperreal.com Subject: Re: How about mixing BindAddress and VirtualHost??? In-reply-to: sanders's message of Fri, 02 Feb 1996 23:07:01 -0600. <199602030507.XAA22996@austin.bsdi.com> X-uri: http://www.zyzzyva.com/ Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Date: Sat, 03 Feb 1996 11:20:51 -0600 From: Randy Terbush Sender: owner-new-httpd@apache.org Precedence: bulk Reply-To: new-httpd@hyperreal.com > Cliff Skolnick writes: > > Well, BindAddress was ment for people that wanted a different > > sever/servers running for addresses. They could be different UID and > > stuff that virtual host directives can't provide. > > I understand how and why Apache got into the current situation, > but the fact remains that if you want six VirtualHosts on one uid > and another six on another uid then you'll have to run seven server > instances instead of two. Ideally, specifying uid/gid for the VirtualHost would be the best solution IMO. One lead parent process forking children under specified UID. This would solve the SetUID debate. This would also have the added side effect of allowing CPU usage accounting on a per UID basis.